The prevalence of OSA is relatively high among people with hypertension (12, 13), type 2 diabetes (14), hyperlipidemia (14), COPD (15), and cognitive decline (12, 16–18). This high prevalence has also been found in patients with lung cancer. Cabezas et al. evaluated 60 patients diagnosed with lung cancer using Home Sleep Apnea Testing and showed that the median apnea-hypopnea index (AHI) was 15.2 (6.4–31.2), the percentage of time spent with saturation less than 90% of the total sleep time (T90%) was a median of 11%, and 50% of patients suffered from moderate-to-severe (AHI ≥15/h) OSA (7). Similar to this study, Bhaisare et al. used polysomnography (PSG) to assess sleep in 30 patients newly diagnosed with lung cancer (29 non-small cell lung cancers [NSCLCs] and 1 small cell lung cancer [SCLC]), and the patients had a mean AHI of 12.01 ± 15.52 events/h and a mean minimum oxygen saturation of 84.53%. Patients that met the diagnostic criteria for OSA were 17 (57%), with 26.6% experiencing mild and 29.9% moderate-severe OSA (19). Another clinical cohort study from Asia confirmed that the prevalence of OSA in patients with lung cancer was 57%, with moderate-severe OSA accounting for 27% of the total, especially in patients with SCLC, an elevated AHI, and oxygen desaturation index (ODI), indicating that hypoxia may be more severe (20). Liu et al. retrospectively analyzed the data of 410 patients with newly diagnosed lung cancer, and a total of 128 patients (31.2%) had OSA, a slightly lower prevalence compared to that reported by other studies (21).

The conclusions are relatively consistent from the information provided in the few available studies. In comparison to the general population, patients with lung cancer are more likely to have coexisting OSA. The diagnosis of OSA may be overlooked in the lung cancer population. Whether this neglect has clinically significant consequences relating to prognosis will be analyzed.

OSA has been demonstrated to increase the incidence of cancer, particularly prostate, breast, and kidney cancers (22). Justeau et al. demonstrated that T90%, a parameter that measures the degree of hypoxia in OSA but not its severity, increased the incidence of cancer by approximately 30% (T90% ≥13 vs. <0.01, hazard ratio [HR] 1.33, 95% confidence interval [CI] 1.05–1.68) (23). This finding was recently confirmed by another study in which nocturnal hypoxemia associated with OSA was found to be an independent risk factor for increased cancer prevalence after fully adjusting for demographic confounders, such as age, sex, and body mass index (BMI) (24).

Simultaneously, scholars have conducted studies on whether OSA has an impact on the incidence of lung cancer. In one of the earliest studies, Gozal et al. demonstrated that OSA did not cause a statistically significant increase in the risk of developing lung cancer (HR 1.02, 95% CI 0.99–1.06) and did not correlate with the progression of lung cancer (25). Subsequently, following veterans for 7 years, Jara et al. demonstrated that OSA increased the incidence of lung cancer by 32% (HR 1.32, 95% CI 1.27–1.38) (26). Similar findings were obtained in another prospective study conducted during the same period (HR 1.52, 95% CI 1.07–2.17). When only non-smokers were included, the incidence of lung cancer in patients with OSA was approximately 3-fold that in patients with non-OSA (HR 2.96, 95% CI 1.42–6.18). However, it should be noted that for this study, an OSA diagnosis was obtained by questionnaire report rather than sleep monitoring (27). Kendzerska et al. also confirmed that severe, but not mild-moderate, OSA contributes to risk factors for lung carcinogenesis (HR 1.34, 95% CI 1.00–1.80). They found that indicators of hypoxia associated with OSA, AHI ≥28, and mean SaO₂ <93.4% were also independent risk factors for lung cancer incidence (28). In a recent meta-analysis (comprising a total of 4,885,518 patients) pooling the above four studies on lung cancer incidence, OSA increased the risk of lung carcinogenesis by 25% (HR 1.25, 95% CI 1.02–1.53), and a further increase was noted if the follow-up time was specified to be at least 5 years or more (HR 1.32, 95% CI 1.27–1.53) (6). Another meta-analysis that included 12 studies (22) showed that, although the prevalence of lung cancer in the general population (0.023%) is slightly lower than the prevalence of prostate cancer (1.1%), OSA prevalence in lung cancer was higher (0.5%) (29). Seijo et al. demonstrated that the severity of OSA (mild, moderate, or severe) was unrelated to the prevalence of lung cancer after adjusting for age and sex. Instead, sleep parameters such as AHI (odds ratio [OR] 1.382, 95% CI 1.015–1.882) and T90% (odds ratio (OR) 1.467, 95% CI 1.121–1.920) were independent risk factors for the development of lung cancer (30). Similarly, a study by Justeau et al. confirmed that neither mild-moderate nor severe OSA increased the incidence of lung cancer, yet T90%, a parameter reflecting the degree of hypoxia, acted as an independent risk factor contributing to the development of lung carcinoma (T90% ≥13 vs. T90% <0.01; HR 2.14, 95% CI 1.01–4.54) (23).

Another clinical cohort study from China could not replicate these studies. Xiong et al. found no significant association between OSA, AHI, T90%, and the incidence of lung carcinoma in 3,786 hospitalized patients with OSA who were followed up for an average of 9 years (31). In 2019, Brenner et al. concluded a study comprising 5,243 patients with suspected OSA (mean age: 51 ± 13.1 years) with up to 5.9 years of follow-up. During this period, altogether, 265 patients were diagnosed with malignant tumors, including 14 with lung cancer (incidence rate of approximately 0.045%), and this study also failed to confirm a link between OSA and the incidence of lung carcinoma (32). Similarly, in a recently completed study, Marriott et al. demonstrated that neither AHI, a parameter reflecting the severity of OSA, nor T90%, a parameter capturing the magnitude of nocturnal intermittent hypoxia, was associated with lung cancer incidence (AHI >30 vs. AHI <5, HR 0.80, 95% CI 0.51–1.26; T90% ≥2.2 vs. T90% <0.1, HR 0.96, 95% CI 0.67–1.38) (24).

Contrary to the findings of the above studies, OSA appeared to be a protective factor for lung carcinogenesis (HR 0.87, 95% CI 0.82–0.93), as confirmed in another long-term follow-up of up to 6 years in an Asian population. In particular, for male patients, OSA was revealed to be a protective factor against the development of lung cancer (HR 0.84, 95% CI 0.78–0.90) (33). Interestingly, similar findings were obtained by Sillah et al. They anticipated 175 cases of lung cancer to occur in 34,402 patients with OSA by age–sex standardized cancer incidence ratios (SIRs); however, only 115 patients eventually presented with lung cancer (SIR 0.66, 95% CI 0.54–0.79) (34).

Overall, 7 of the 13 studies mentioned above confirmed that OSA may increase the incidence of lung cancer, including two meta-analyses with a high level of evidence; two studies found a negative association. However, we cannot conclude these studies based on their heterogeneity, including differences in study design, follow-up time, inclusion population, and reference indicators. The first aspect to consider is the variability between study designs. All but one study (30) had relatively large sample sizes; however, only one was a prospective study (27), and the diagnosis of OSA was not exclusively confirmed by sleep monitoring, even though Huang et al. examined PSG data in a sample of 108 individuals, 98% of whom met the criteria for the diagnosis of OSA. These inconsistencies can lead to a decrease in the reliability of the conclusions of the studies.

Meanwhile, among the studies, the duration of follow-up ranged from 1.9 to 9.1 years. OSA is a chronic disease, and its effects on human health can be slow and long-term (35, 36); an insufficient length of follow-up results in a reduced number of positive events, thus affecting the accuracy of the final results. In those studies that have demonstrated that OSA can raise the risk of lung carcinogenesis (26–28), the average duration of follow-up was more than 7 years. The second drawback is that although the studies, including two meta-analyses, have demonstrated that OSA increases the risk of lung cancer, only Seijo et al. (30) and Kendzerska et al. (28) have demonstrated a causal relationship between OSA-associated hypoxia (AHI, ODI, mean SaO₂) and the elevated incidence of lung cancer. It is also noteworthy that mean SaO₂ is more reflective of the degree of nocturnal hypoxia rather than being a marker specific to IH. Considering that many confounding factors accompanying OSA, such as aging, smoking, and obesity, also increase the risk of lung cancer, it would be highly arbitrary to generalize that IH because of OSA increases the incidence of lung cancer. Finally, these studies did not consider a pathophysiological feature of OSA, namely SF. The duration of sleep is linked to the progression of cancer, and sleep deprivation increases the risk of its development (37, 38). For OSA, SF is an objective reflection of sleep deprivation, and SF interacts with IH to create a vicious cycle. For example, IH can lead to microarousals and, therefore, to SF. In turn, decreased arousal thresholds and SF lead to apneic events and IH (39). Pathophysiologically, SF is still distinct from sleep deprivation and shortened sleep duration alone. There is a lack of consistent metrics to quantify SF, which may limit studies on the correlation between SF and lung cancer. However, it would be interesting if multiple parameters in PSG that reflect arousal and SF (e.g., microarousal index, electroencephalogram (EEG)-based sleep staging conversion, and electroencephalographic energy changes) were to be included in future studies.

OSA has a high prevalence among patients with lung cancer. AHI >30/h has been shown to result in a higher risk of all-cause mortality (40), while lung cancer is also a common cause of death (41); however, it is unclear whether the combination of the two would further increase the risk of mortality. Prognostic studies addressing concurrent OSA and lung cancer are limited. The results of a follow-up study of 45 patients with lung cancer (24 combined with OSA) confirmed that in the OSA group, a total of 10 deaths and 9 recurrences or metastases occurred, with a significantly greater overall rate of deterioration (death + recurrence or metastasis) than that in the non-OSA subgroup (5 deaths and 5 recurrences or metastases) (42). Liu et al. conducted a study of 44 patients with NSCLC (OSA: 22, non-OSA: 22) with a 2-year follow-up. The survival rates of patients with OSA were lower than those without sleep apnea (1-year: 90.9% vs. 95.5%, 2-year: 65.8% vs. 71.6%), although no significant difference was found between the two groups based on the log-rank test. Vascular endothelial growth factor (VEGF) is a proangiogenic cytokine that contributes to the development of solid tumors by promoting tumor angiogenesis (43). Currently, various anti-VEGF drugs, such as bevacizumab, are being utilized for tumor treatment. After measuring the plasma levels of VEGF, the study found that the expression of VEGF was significantly higher in the OSA+NSCLC group (44). A study by Huang et al. confirmed that in intermediate and advanced lung cancer (stages III and IV), severe OSA increased the risk of death, with 80% of patients with severe OSA+ lung cancer dying within three years. Patients with mild-moderate OSA had longer PFS and OS than the severe OSA group (8). Data from the above three studies were included in a recent meta-analysis (a total of 112 patients with lung cancer, including 67 with coexistent OSA), and it appeared that OSA did not increase the risk of death from lung cancer (OR 2.005, 95% CI 0.703–5.715) (37).

To date, no studies have confirmed an increased risk of mortality from lung cancer caused by OSA. However, drawing such a conclusion might be hasty. It could even misdirect the clinical management of this group of patients, given the high prevalence of OSA in patients with lung cancer. Our rationale is as follows. First, the studies included a very small population, with sample sizes totaling less than 200. Second, factors associated with the prognosis of lung cancer include clinical stage and pathologic and molecular type, which are elements that need to be clarified in the clinical management of lung cancer. Third, there are many indicators for evaluating the prognosis of lung carcinoma, such as disease-free survival (DFS), PFS, OS, and 3- or 5-year survival rates (45, 46). Study designs should not be one-sided based on a single outcome metric. Finally, among the above studies, except for one with a follow-up duration of approximately 5 years (8), the median follow-up time in the other two was less than 2 years (42, 44), and the shorter follow-up time could affect the reliability of the study conclusions. By including a large sample of patients with lung cancer, stratifying the design according to lung cancer type, referring to several outcome indicators, and using the last patient death as the follow-up endpoint, the conclusions of the studies can be made more convincing. Unfortunately, none of the current studies have addressed these issues. Accordingly, we cannot yet conclude this issue, i.e., whether OSA affects the prognosis of patients with lung cancer.

Epidemiologic studies addressing the correlation between lung cancer and OSA have reported mixed results ( Table 1 ). Limited findings indicate that OSA is relatively common in patients with lung cancer. OSA is also associated with an increased incidence of lung cancer. To date, there is a lack of direct evidence that OSA raises the mortality of lung cancer. Based on the above analysis of the study limitations, we believe that there is an urgent need to further improve the quality of studies, e.g., by conducting prospective, large-sample, multicenter studies and setting the lung carcinoma incidence and prognosis to the primary outcomes, to draw more convincing conclusions and guide clinical practice.

The tumor immune microenvironment (TIME) comprises immunosuppressive molecules, matrix components, inhibitory immune cells, and cytokines. Its primary function is attenuating the anti-tumor immune response, sustaining cell proliferation, preventing cell apoptosis, maintaining an immunosuppressive milieu, and facilitating angiogenesis (49, 50). Xie et al. (51) discovered that the responses of various immune cell types in the peripheral blood of patients with OSA exhibited dissimilarities and were positively associated with the severity of hypoxemia. Studies have demonstrated that IH exerts its immunosuppressive effects through various hypoxia-inducible factor-1a (HIF-1α)-related signaling pathways: inducing remodeling of the HIF-1α metabolic pathway and enrichment of lactate in the TIME; impeding T-cell proliferation, tumor invasion, and cytokine production; and augmenting myeloid-derived suppressor cell (MDSC) numbers while inhibiting natural killer (NK) cell and CD8⁺T cell activity (52). Simultaneously, IH-induced tumor cells release interleukin 10 and promote M2-type differentiation of tumor-associated macrophages (TAMs) (53), which exhibit immunosuppressive properties. This further fosters the accumulation of other immunosuppressive populations, such as MDSCs, granulocytes, and regulatory T cells (Tregs) (54). Akbarpour et al. constructed a murine model of NSCLC in vitro that mimicked OSA, which demonstrated that IH and SF reduced tumor killing by cytotoxic T lymphocytes (CTLs) by significantly decreasing the secretion of granzyme B by CTLs (55). The alteration in immune function allows cancer stem cells (CSCs) to undergo immune evasion and maintain their self-renewal capacity, the growth and progression of lung cancer (55).

TGF-β plays a pivotal role in the initiation and progression of tumorigenesis. The classical and non-classical signaling pathways mediated by TGF-β participate in tumor cell growth, invasion, and metastasis and play a crucial role in regulating diverse immune cell types within the tumor microenvironment. TGF-β has emerged as a central immune regulator in the tumor microenvironment, and its combination with immunotherapy can exert significant anti-tumor effects (56). An analysis of the phenotype and immune response activity of 29 healthy volunteers and 32 patients with continuous positive airway pressure (CPAP)-treated OSA revealed that glycoprotein-A repetitions predominant protein monocytes from untreated patients with OSA inhibit NK cells through the release of TGF-β; however, reoxygenation eventually restores their altered phenotype and cytotoxicity (57). Subsequent studies have demonstrated that IH can enhance lung cancer cell migration by upregulating the TGF-β signal, increasing the activation and proportion of lung cancer-associated fibroblasts (CAFs), thereby promoting lung cancer progression (58).

The programmed cell death protein 1/programmed cell death-ligand 1(PD-1/PD-L1) pathway is an important signaling pathway for the body to negatively feedback suppress immune responses and maintain autoimmune homeostasis (59, 60). In the tumor microenvironment, the over-activated PD-1/PD-L1 pathway inhibits the body’s immune surveillance and immune clearance effects on tumor cells (61). In a murine model of sleep apnea (48), PD-L1 expression levels were significantly increased in the intermittent hypoxia group, and HIF-1α levels were strongly correlated with PD-L1 expression. In a multicenter observational study, 360 patients with cutaneous melanoma were enrolled, and serum soluble programmed death ligand 1 (sPD-L1) levels were measured, which showed that sPD-L1 levels were higher in patients with severe OSA than in those with non-severe OSA and those with non-OSA (62). Carolina et al. showed that IH-induced PD-L1/PD-1 overexpression in patients with severe OSA reduced autologous T cell proliferation and cytotoxic activity of CD8⁺T cells, increased recruitment of myeloid-derived suppressor cells, and showed an increase in PD-L1 on monocytes and PD-1 on CD8⁺T cells in patients with OSA (63). Theoretically, the higher the PD-L1 expression, the stronger the tumor cell immunosuppression. IH associated with OSA promotes the expression of PD-L1 in tumor cells, and by blocking the binding between PD-1 and PD-L1, PD-1/PD-L1 inhibitors enable immune cells to maintain their activity and restore their recognition and killing of tumor cells. Therefore, we speculate that PD-1/PD-L1 inhibitors may have a significant effect on the treatment of lung cancer patients with OSA, which might be an interesting area of research to be further confirmed in the future.

TAMs, MDSCs, and Tregs are involved in tumorigenesis, invasion, and metastasis. TAMs can be differentiated from MDSCs and are involved in constructing tumor stroma. MDSCs themselves can suppress T cell responses. Under various chemokines and cytokines, monocytes are recruited to the periphery of lung cancer cells and differentiate into TAMs. Under hypoxic conditions, TAMs secrete mitogenic factors and immunosuppressive agents that promote the progression of lung cancer (64). The phenotypes that make up human TAMs include two main types, M1 and M2. M1-TAMs play a role in inhibiting lung cancer progression by killing tumor cells.

Conversely, M2-TAMs are promoters for tumorigenesis and progression. They inhibit cellular immune responses and the activity of natural killer cells (NK cells), releasing inflammatory factors and protein hydrolases. They are considered important immunosuppressive cells in the tumor microenvironment (65). The polarization process (M1 to M2) of TAMs, i.e., the decrease in M1/M2 values, has also been proposed to predict the poor prognosis of lung cancer (66). IH associated with OSA can further promote the growth of lung cancer tissues by promoting the differentiation and polarization of TAMs (M1 to M2), as well as increasing the invasiveness to the surrounding tissues and distant metastasis of lung cancer cells. Almendros et al. constructed an IH model mimicking OSA and implanted a lung epithelial tumor into mice. They found lung carcinoma cells in the IH group were more invasive than those without hypoxia after 28 days. Concurrently, compared to the control group, there was a 2.6-fold increase in TAMs around the tumor tissue (P=0.02), with a simultaneous increase in Tregs and MDSCs (2.2-fold and 3.7-fold, respectively). Applying TAMs isolated from IH mice to in vitro cultures significantly increased tumor cell proliferation, migration, and invasiveness ( Figure 1A ) (67). A similar study also confirmed that IH-exposed mice had significantly higher lung cancer tissue weights (1.363 ± 0.143 g) than hypoxia-free mice (0.677 ± 0.097 g), with a 2.2-fold increase in IH-induced TAM infiltration, which further confirmed that prolonged exposure to IH induces adipose tissues to recruit adipose tissue stem cells (ASCs) and pro-inflammatory cells, leading to the migration of adipose tissue macrophages and ASCs to tumors, promoting tumor tissue growth and enhancing its invasiveness (70). The effect of sex on the relationship between IH and lung cancer was first explored by Torres et al. by mimicking postmenopausal mice (oophorectomized). They demonstrated that, compared to mice exposed to CH, IH promoted greater lung cancer tissue growth in ovariectomized mice. Although not statistically significant (P=0.08), this study also demonstrated that IH promotes TAM aggregation around tumor tissues and facilitates the conversion of TAM phenotype from M1 to M2 ( Figure 1B ). This study confirmed that IH associated with OSA can facilitate the progression of lung carcinoma (68). Recently, a study confirmed that TAMs and PD-L1 expression were upregulated in patients with OSA with lung adenocarcinoma. OSA-related IH activated HIF-1α and its associated signaling pathways, enhanced the expression of PD-L1 in tumor cells, and increased the activity of TAMs while decreasing cytotoxic T cell activity. Lung cancer growth was promoted through the tumor-associated immune response mediated by IH (71). Campillo et al. constructed an IH+NSCLC model in vivo and in vitro (69). The exposure of IH promoted the proliferation of lung cancer by upregulating the number of TAMs as well as mediating the polarization of the phenotype of TAMs (the ratio of M2/M1 was increased 2-fold in the IH group compared to that in the control group). Cyclooxygenase-2 and its downstream secreted prostaglandin E2 (PGE2) were shown to be important mediators in regulating this process. PGE2 expression was 1.5-fold higher in the IH+NSCLC group ( Figure 1C ). This study confirmed that OSA-associated IH exacerbated lung cancer progression by activating the inflammatory response pathway (COX-2/PGE2) and promoting the proliferation of TAMs (69).

Since the proposal of CSCs by Mackillop et al. in 1983 (72), they have been acknowledged as a highly tumorigenic subpopulation capable of facilitating tumor growth, drug resistance, recurrence, and metastasis through the process of epithelial-mesenchymal transition (EMT) reprogramming (73, 74).

The study by Gu et al. conclusion that IH promotes the differentiation, proliferation, invasion, and migration of CSCs through HIF-1α and can simultaneously induce drug resistance. This study also identified another important molecule: endothelial cell-specific molecule-1 (ESM1), and using an ESM1 antagonist, they reversed the negative effects of IH on a murine model of NSCLC ( Figure 2A ) (75). In vitro and vivo, Hao et al. demonstrated that IH promoted tumor proliferation, invasion, and distant metastasis and identified a series of mechanisms by which IH promotes the progression of NSCLC through the upregulation of CSC expression. In vivo and in vitro, IH promoted the aggregation of mitochondrial reactive oxygen species (mtROS), mediated oxidative stress, and induced the generation of more CSCs by upregulating the expression of transcription factor BTB and CNC homology 1 (Bach1). This response mediated by IH disappeared after gene silencing of Bach1 (Figure 2B) (76). Meanwhile, by measuring the levels of HIF-1α and ATPase family AAA structural domain-containing protein 2 (ATAD2)-associated genetic products in A549, H460, H1299, and SPC-A1 cells, which are the representative NSCLC cells, the levels of ATAD2-related mRNA and protein levels were found to be significantly elevated in the IH+NSCLC model. Knockdown of ATAD2 significantly inhibited the invasion and migration of lung carcinoma cells. The production of mtROS and the number and activity of CSCs were also reduced. This study confirmed that IH regulates the interaction between mtROS and CSCs by mediating HIF-1α/ATAD2, promoting lung cancer progression (77).

Continuous neovascularization guarantees the tumor tissue’s thirst for nutrients, necessary to maintain lung carcinoma cells’ proliferation, infiltration, and distant metastasis. Vascular endothelial cells are the cornerstone in this process, and VEGF promotes lung cancer carcinogenesis and progression by inducing the differentiation and maturation of endothelial cells (78, 79).

Zhang et al. investigated the expression of serum VEGF and endothelin-1 in lung cancer associated with IH. Their results showed that the volume and weight of tumors exposed to IH were significantly higher than those of tumors under normal oxygen conditions (3974.50 ± 1748.20 vs. 2268.54 ± 1874.48 mm² and 5.76 ± 1.48 vs. 4.49 ± 0.96 g, respectively). VEGF and its mRNA levels were also significantly increased in serum as well as tumor tissues of mice. This research also demonstrated that IH promoted the proliferation and differentiation of endothelial cells by upregulating the expression of VEGF and endothelin-1, which accelerated the progression of lung cancer by inducing neovascularization in tumors ( Figure 3A ) (80). The same finding was replicated in another study during the same period. Kang et al. confirmed the role of VEGF and endothelial cells in promoting IH-mediated lung cancer progression. The expression level of VEGF was markedly higher in the IH+lung adenocarcinoma mice model than in the normoxia+lung adenocarcinoma group (308.1 ± 104.3 vs. 172.0 ± 90.6 pg/mL). Nuclear factor-erythroid 2-related factor 2 (Nrf2) and Beta-catenin (β-catenin), but not HIF-1α, appear to be the key intermediary mediators promoting VEGF expression ( Figure 3B ) (81).

Serum midkine (MDK), as a lymphangiogenesis-related biomarker, its overexpression has been identified in patients with NSCLC as a potential therapeutic target for its development, metastasis, and angiogenesis (82, 83). Lymphangiogenesis, in addition to angiogenesis, mediates tumor vascularization, which meets the oxygen and nutrient requirements for tumor growth and metastasis (84). Recent studies have shown that individuals at high risk for lung cancer with moderate-to-severe OSA have elevated expression of MDK (non-OSA: 1536 pg/mL, 95% CI 840–2360 pg/mL; moderate-to-severe OSA: 5902 pg/mL, 95% CI 816–8337 pg/mL) (85).

Exosomes are important bioactive factors that can participate in cell communication, immune response, cell differentiation, and tumor invasion and can also be used as carriers for drug therapy (86, 87). Exosomes increase the proliferation of lung cancer by stimulating glucose anaerobic fermentation metabolism and ensuring oxygen supply to lung cancer cells by promoting neovascularization during lung cancer carcinogenesis and progression. Concurrently, exosomes can induce EMT and deliver regulatory factors between cells, thus promoting distant tumor metastasis (88).

Almendros et al. extracted exosomes from blood samples from 10 adult patients with OSA in fractions (before CPAP treatment and after 6 weeks of treatment) and confirmed that exosomes before CPAP treatment significantly enhanced the differentiation, maturation and migration of adenocarcinoma cells, which was validated in an animal model. In parallel, this study also identified several miRNA sequences (such as mmu-miR-92a-3p and mmu-miR-709) associated with exosomes ( Figure 4A ) (89). Subsequently, a study by Liu et al. also confirmed that exosomes derived from lung cancer cells under IH-exposed conditions upregulated the expression of PD-L1 on the surface of macrophages surrounding tumor tissues and promoted lung cancer progression ( Figure 4B ). Compared with patients with lung cancer without OSA, the percentage counts of monocytes positively expressing PD-L1 were markedly higher in serum samples from patients with lung carcinoma with concomitant OSA (95.3% vs. 86.0%), and this result was positively related to the parameters AHI and ODI, which reflected the severity of OSA (r=0.419 and 0.387, P<0.05, respectively) (90).

The metabolism of normal tissues and cells results from a combination of multiple gene regulation processes. Proto-oncogenes promote the process of cell mitosis and developmental maturation by regulating the cell cycle. Oncogenes mainly prevent excessive cell proliferation. The dynamic balance between the two maintains the normal metabolism of the body. The carcinogenesis and progression of lung carcinoma is the adverse consequence of dysregulation of proto-oncogenes/oncogenes (91–94).

Genetic material is equally vital in regulating the carcinogenesis and progression of OSA-related IH and lung carcinoma. Cortese et al. studied NSCLC mice exposed to IH. They confirmed a significant increase in the amount of cirDNA compared to NSCLC mice in normoxic environments by measuring cirDNA levels in plasma (difference between the two: 510.62, P=0.015). Meanwhile, with the increase in the amount of cirDNA, the size and weight of the tumor increased (R^{2 =} 0.580 and 0.765, respectively, P<0.05). These cirDNAs may carry specific epigenetic modifications for information transfer between OSA and lung cancer (95). RNA is an important carrier of genetic information, controlling cell growth and differentiation by conducting protein synthesis, transmitting genetic information, and regulating transcription and translation (96). Chao et al. constructed human lung adenocarcinoma cell lines exposed to IH in vitro and inhibited the proliferation and invasiveness of IH-exposed tumor cells by knocking down the demethylation enzyme ALKB homologous protein 5 (ALKBH5). This process may be mediated by downregulation of the level of n6-methyladenosine (m6A)-modified mRNAs. This finding was confirmed in vivo and in vitro (97). MicroRNAs are a class of small RNAs about 20–24 nucleotides in length that do not encode proteins but can regulate the translation of mRNAs. Li et al. identified a specific fragment of genetic material, microRNA-320b. In patients with lung cancer who also experienced OSA, IH induced a decrease in the expression of microRNA-320b, with downstream expression of ubiquitin-specific peptidase 37 (USP37) upregulated secondary to IH and USP37 promoting lung cancer progression by mediating the deubiquitination of Cdc10-dependent transcript 1 (CDT1) (98). Public databases and big data platforms have recently provided us with rich research resources. Zhao et al. analyzed 734 patients (230 adenocarcinomas and 504 squamous carcinomas) with NSCLC accompanied by OSA by combining genetic databases. This study confirmed that the metabolic pathway-related genes HK2 and GBE1 were significantly altered in lung cancer cells under IH conditions. HK2 and GBE1 are located downstream of the HIF-1α signaling pathway, and their expression was positively correlated with tumor size, primary tumor stage T, and tumor lymph node stage N. Meanwhile, higher expression of HK2 and GBE1 has also been shown to affect OS in lung adenocarcinoma but not squamous lung cancer (99). From the Gene Expression Omnibus and The Cancer Genome Atlas, Wang et al. used a weighted gene co-expression network analysis screening method to identify four key genes that are shared with OSA and lung cancer, namely modulator of apoptosis 1 (MOAP1), chromobox 7 (CBX7), platelet-derived growth factor subunit B (PDGFB), and mitogen-activated protein kinase 3 (MAP2K3). In this study, a model was constructed to predict the prognosis of patients with lung cancer using an innovative machine-learning approach. Although the area under the receiver operating characteristic curve for this model was 0.707, indicating that the model was moderately effective, this study pioneered the combination of artificial intelligence and OSA with lung cancer, providing ideas for future research (100). Qi et al. also applied a machine learning approach to the study of OSA and NSCLC by screening for two OSA-associated target genes (EXO1 and KRT6A) affecting lung adenocarcinoma treatment and prognosis and obtaining three sets of clusters after clustering and analyzing the relevant populations. Patients in Cluster 1 were the most sensitive to conventional chemotherapeutic agents but also had the worst prognosis. Patients in Cluster 2 had a poor response to immunotherapy, which could be attributed to T cell depletion because of the chronic inflammatory response associated with long-term IH. In contrast, those in Cluster 3 could maximize the benefits of treatment with immune checkpoint inhibitors (101). This finding provides a new biomarker for predicting lung adenocarcinoma prognosis, although the model requires external validation.

The carcinogenesis and progression of lung cancer involve various factors, including genetic mutations, immunosuppression and evasion, and alterations in the immune microenvironment ( Figure 5 ). The two core pathophysiological mechanisms of OSA, IH and SF, lead to secondary oxidative stress, chronic inflammatory response, and dysfunction of the neuro-endocrine and immune functions, closely associated with cancer (102). IH can cause changes in cells, factors, and signaling pathways in the TIME, such as remodeling HIF-a1 and TGF-β signaling pathways, activating TAMs, CAFs, and other cells, and promoting PD-1 expression. Meanwhile, scholars have attempted to establish a bridge connecting the pathophysiological link between OSA and lung cancer and have confirmed that IH, which is associated with OSA, results in the overexpression of some signaling molecules that promote cell proliferation by activating the HIF-1α signaling pathway, which further activates downstream signaling pathways (including inflammation, cell proliferation, and neovascularization). CSCs influence and modify the tumor microenvironment, and vascular endothelial cells differentiate and mature under VEGF stimulation to participate in neovascularization, supplying nutrients required for the growth of lung cancer cells. Although HIF-1α responds differently to IH, sustained hypoxia, and acute hypoxia, IH activates HIF-1α more rapidly than sustained hypoxia (103).

Meanwhile, IH and SF can also promote the recruitment of MDSCs to lung cancer tissues, induce the polarization of TAMs, and inhibit the function of Tregs by regulating the immune system. Exosomes are emerging bioactive factors with multiple biological functions closely related to the progression of lung cancer and therapeutic resistance by transporting nucleic acids and proteins, transmitting biological information, and participating in tumor cell migration and differentiation, immune response, and regulation (104). Long-term exposure to IH induces the activation of proto-oncogenes, and the inactivation of oncogenes, and the body eventually loses the regulation of normal cell proliferation/apoptosis. Many of the above factors combine to contribute to the progression of lung cancer, and these factors may also be promising therapeutic targets.

The entire process of lung cancer progression includes carcinogenesis, tumor tissue proliferation, local invasion, and distant metastasis. The studies mentioned above explain the role played by IH in promoting local invasion and tumor tissue proliferation of lung cancer cells. However, one current limitation is that no study has been able to confirm that IH activates lung carcinogenesis, i.e., that there is a direct causal relationship between IH and lung cancer genesis. In future studies, it is necessary to determine if IH has direct tumorigenicity for lung cancer. Second, a lack of standardized animal models adequately explains the heterogeneity between different studies. In addition, as another important pathophysiological mechanism of OSA, SF may be more complex than IH (105). Unfortunately, to date, only one study has explored the effect of SF on lung cancer (55), and even in that study, SF was not specifically quantified. Finally, IH associated with OSA is also complex, with differences in the duration, frequency, and depth of hypoxia leading to large differences in the degree of hypoxia between patients with a similar AHI. Current in vivo animal models or in vitro cellular models do not simulate the most realistic state of OSA, which underlies the difficulty of the current work.